Last Build Date: Wed, 07 Sep 2016 22:24:09 +0000
Wed, 07 Sep 2016 22:24:09 +0000I have a question. How does one account for delisted or failing companies in this or any backtest? Thanks.
Thu, 21 Jul 2016 16:53:18 +0000do you use nopat (i.e., operating cf x (1- tax) as the cf metric? Also, do you break out capex into maintenance and growth; realize that this is very difficult to do. I think i would agree with you that I would not subtract dividends from cf as a measure of value as it would apprear that the test is to measure the cf generating capacity of the business.
Fri, 08 Jul 2016 14:16:20 +0000We have not had much success testing ROI. We however did find another quality ratio that works. It was developed by Professor Novy-Marx who defined a quality company as one that had a high gross income ratio calculated by dividing gross profits by total assets. He defined gross profit as sales minus cost of sales and assets simply total assets as shown in the company’s balance sheet (current assets + fixed assets). You can see the full back test here: www.quant-investing.com/blogs/general/2015/01/28/have-you-been-using-the-wrong-quality-ratio
Fri, 08 Jul 2016 14:02:28 +0000Another interesting back test thanks George. I prefer 50 million market cap because I invest my own money and would rather look for ideas where there are less interest by analysts and large investors. What may be a better selection criteria is trading volume. For example $100,000 30 day average which should be more than enough for small investors. We tested TTM FCF to EV and five year average FCF to EV on European companies over the 12 year period 13 June 1999 to 13 June 2011. TTM FCF worked a better than five year average FCF but we found that if you combine TTM FCF with other ratios you get much better returns. TTM FCF combined with 12 month share price momentum gave the best result of +755% You can read more about the back test here: http://www.quant-investing.com/strategies/free-cash-flow-yield-and-5-year-average-free-cash-flow-yield
Sat, 21 May 2016 05:45:55 +0000"either the trader wins and earns as high as 80% returns or he loses the trade and also his initial investment. " sounds like a pretty bad deal, huh?
Sat, 21 May 2016 05:42:36 +0000Hi, thank you for sharing! "Finance is something most people don’t start learning about until high school or even college" sadly true kind regards valuetradeblog
Wed, 18 May 2016 04:22:05 +0000Really helpful piece of content you shared here! Calculating an ROI is never easy and I have to say you made it easy to understand and nailed it the right way. Cheers
Tue, 10 May 2016 14:45:57 +0000George we tested Dividend yield, Sharedolder Yield (James O’Shaughnessy) and Shareholder Yield (Mebane Faber) in Europe over a 12 year period from June 2001 to May 2013. Results (CAGR): Market 5.65% Dividend yield 9.95% Sharedolder Yield (James O’Shaughnessy) 9.94% Shareholder Yield (Mebane Faber) 10.04% So not much of a difference. What worries me with dividend strategies is taxes on dividends which are not in most back tests (mine included). Here in Germany I pay around 36% in taxes. That is a huge return destroyer! You can read the whole backtest here: http://www.quant-investing.com/blogs/general/2014/11/18/dividend-investor---this-ratio-beats-dividend-yield @Manik We tested value and momentum with VERY good results here: http://www.quant-investing.com/blogs/general/2015/01/20/1374.2-return-from-neglected-value-and-momentum-companies @Murik After reading David Dreman's book Contrarian Investment Strategies a long time ago I became VERY careful with using forecasts or estimates. His research showed they are basically useless. His book is a great read!
Sun, 01 May 2016 16:46:39 +0000How about growing current and next Fiscal Year EPS and Sales estimates?
Sat, 30 Apr 2016 01:43:51 +0000An asset growth backtest does sound interesting. I plan on doing multi-factor models once I start running out of single factor ideas.
Fri, 29 Apr 2016 16:14:15 +0000Asset growth backtest could be interesting. What about multi-factor models? How would you combine momentum and value?
Thu, 28 Apr 2016 08:42:20 +0000We tested FCF Yield and 5 year average FCF Yield in Europe over the 12 year period 13 June 1999 to 13 June 2011. 12 months FCF yield looked a lot better than 5 year average FCF yield. We also tested 12m FCF yield against a lot of other indicators and found 12 months momentum the best increasing returns by 430% Odd was that Price to book as a second factor also did well but this may have something to do with the test period with three large market declines after which we have found price to book does really well. You can see all the results here: http://www.quant-investing.com/strategies/free-cash-flow-yield-and-5-year-average-free-cash-flow-yield
Tue, 26 Apr 2016 15:18:04 +0000George all good points. I looked at this problem from a screening perspective to get rid of low quality net net ideas. My best ideas at the moment is to look for net nets with a healthy > 5 Piotroski F-Score or look for companies with positive free cash flow (cash from operations - capex) or both Here is an article I wrote about it: http://www.quant-investing.com/blogs/general/2014/10/28/time-to-start-looking-for-net-nets We have not tested the strategy though. Not directly related but here is an interesting paper I found recently: http://www.quant-investing.com/blogs/general/2016/02/02/beat-the-market-with-small-companies-but-you-have-to-do-this I also suspect that net net and low price to book strategies work best after a substantial market decline. Here is great work by Jim O'Shaughnessy http://jimoshaughnessy.tumblr.com/post/103140701394/price-to-book-value-ratios-a-long-term-winner
Thu, 21 Apr 2016 20:54:11 +0000Thanks George - I was referring to using FactSet Estimates or IBES Estimates as the source... not sure if Portfolio123 has the ability to integrate these content sets, but basically FDS and Thomson just go out and aggregate broker consensus estimates (FDS is industry standard).
Thu, 21 Apr 2016 18:09:23 +0000I use it in almost all of my screens and backtests. It does slow them down a bit though, so I usually put it after the first few rules that cut down the size of the remaining universe. I just got so frustrated with the "average" rule, because so many stocks would come through that couldn't be trusted to be liquid. And it's easy to customize for different amounts and time frames... :) That actually came from Portfolio123 when I asked about a minimum instead of an average. They just recently added a LOOPMIN() function that might be easier to use. Or, six of one, half dozen of another.
Thu, 21 Apr 2016 17:06:31 +0000Hi Tim, Thank you for the suggestion to backtest shareholder yield. I'm going to try something similar to Mebane Faber's formula. However, to avoid issues with stock splits, I'm going to define shareholder yield as follows: (DivPaid(0,TTM)+DbtTotPYQ-DbtTotQ+( EqPurch(0,TTM)-EqIssued(0,TTM)))/mktcap What do you think of that definition?
Thu, 21 Apr 2016 15:18:52 +0000Randy, Thank you for sharing that formula for minimum daily trading value. I'm now using something similar for my backtests.
Thu, 21 Apr 2016 15:17:17 +0000Hi Justin, I would look at NTM FCF/Market Cap or NTM FCF/EV if I knew of a consistent source for next twelve month free cash flow estimates. I'm not aware of any source for that data.
Thu, 21 Apr 2016 03:57:32 +0000have you run NTM FCF/MKT VAL or NTM FCF/EV? would be interesting to compare results between ntm and ltm.
Wed, 20 Apr 2016 15:12:34 +0000Love it! Smart to always look at the fundamentals. Looking forward to your next post.